(Bringing it back to the list)
On 02/14/2013 02:54 PM, Khdher Omar wrote:
HI ,
Thanks for response. In fact , I am making tests without usage of cephx
or any other authentication mecanism.
I am looking for a proof of concept with basic boot from volume.
is cephx required for that issue ?
No, cephx is not required. It's just that I noticed it.
Looking at the XML a bit further I don't see a monitor host defined,
there should be something like:
<source protocol='rbd' name='rbd/3b987fba-5153-4e51-b346-cfe20ecd9d27'>
<host name='ip.of.monitor.host' port='6789'/>
</source>
But I think that librbd/librados in this case reads the
/etc/ceph/ceph.conf instead of getting all the arguments through libvirt.
On the compute node, do things links:
$ rbd ls
$ rbd info <image name>
Wido
Omar
------------------------------------------------------------------------
*De :* Wido den Hollander <wido@xxxxxxxx>
*À :* ceph-users@xxxxxxxxxxxxxx
*Envoyé le :* Jeudi 14 février 2013 14h45
*Objet :* Re: RBD boot from volume in OpenStack with Ceph
Hi,
On 02/14/2013 12:44 PM, Khdher Omar wrote:
> The issue about OpenStack related with Ceph. Thus, I was following
> documentation in Ceph website on how integrate Ceph with existing
> OpenStack environment.
> I used Folsom with two nodes. I just configured Ceph with Cinder and
> Glance and I could make RBD as storage backend for Glance. Still with
> volumes that i was looking for to boot from it.
> I proceed afterwards by :
> 1. create a new rbd pool on ceph;
> 2. create new volume [cinder create volume 2 for example];
> 3. Check if the new volume is added to ceph pool created --> success
> 4. Finally boot form volume :
> nova boot --flavor m1.tiny --image precise-ceph --block_device_mapping
> vda=ID_VOL:::0 --security_groups=default boot-from-rbd
> last step failed : Instance created but with Error Status.
> We could check inside the
> /var/lib/nova/instance/instance_id/lilbvirt.xml an intersting section:
> <disk type="network" device="disk">
> <driver name="qemu" type="raw" cache="none"/>
> <source protocol="rbd"
> name="rbd/volume-e5d5f756-7a9d-47b1-a59d-504c9cf582d5"/>
> <target bus="virtio" dev="vda"/>
> </disk>
I'm not so familiar with OpenStack, but looking at the XML from libvirt
I'm missing a "secret".
Are you using cephx? If so, did you define a libvirt secret and
configure the UUID in Nova?
Wido
> basically it seems that the opensstack recognized for RBD using the
> virtio driver.
> We could check the log file 'nova-compute.log' the follwing error line:
> self._set_instance_error_state(context, instance['uuid'])
> TRACE nova.openstack.common.rpc.amqp libvirtError: internal error
> Process exited while reading console log output: char device redirected
> to /dev/pts/6
> TRACE nova.openstack.common.rpc.amqp kvm: -drive
>
file=rbd:rbd/volume-e5d5f756-7a9d-47b1-a59d-504c9cf582d5,if=none,id=drive-virtio-disk1,format=raw,cache=none:
> error connecting
> it seems that the if field shows none.
> Is there any explanation that KVM can't read that volume ?
> Thanks for any response !
>
>
> _______________________________________________
> ceph-users mailing list
> ceph-users@xxxxxxxxxxxxxx <mailto:ceph-users@xxxxxxxxxxxxxx>
> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
>
--
Wido den Hollander
42on B.V.
Phone: +31 (0)20 700 9902
Skype: contact42on
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx <mailto:ceph-users@xxxxxxxxxxxxxx>
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
--
Wido den Hollander
42on B.V.
Phone: +31 (0)20 700 9902
Skype: contact42on
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
